open TELEMAC-MASCARET The mathematically superior suite of solvers

Hi all!

I recently ran a sediment plume case and I have problem with mass balance.
I have two sediment concentration inflow boundaries with:
- C1 = 0.000131 (adimensional, = 0.347 kg/m3)
- C2 = 0.0000189 (adimensional, = 0.05 kg/m3)
Both these concentrations are associated with liquid boundaries with a constant discharge of 1 m3/s.
I have only one sediment class.

My simulation is run during 8h, which means that my total mass input is:
- M1 in = 1440 kg
- M2 in = 9994 kg

When looking at the quantity of sediment leaving through boundaries, I found unexpected high values, higher than the total input mass (I also have some deposition inside the model boundaries)!
- M1 out = 2004 kg
- M2 out = 10183 kg

I attach my Sisyphe steering file which shows the parameter I used.

I don't have a lot of experience with Sisyphe and I don't know if I am doing something wrong and how to improve the mass balance.

I can mention the following:
- I defined the whole model area as non erodible in noerod.f
- I have EQUILIBRIUM INFLOW CONCENTRATION = NO
- liquid boundaries (where sediment inflow is applied) are defined with VELOCITY PROFILE = 4
- sediment inflow boundaries are defined (in sisyphe conlim file) with: 4 5 5 ... 4
- sediment outflow boundaries are defined (in sisyphe conlim file) with: 5 4 4 ... 5
- the simulation log indicates that the max CFL for suspension is approx. 0,6 which I assume is suffienctly low (?)
- the hydrodynamic simulation is stable

I also find difficult to understand the meaning of the final mass balance expressions used in the log file:

*** FINAL BALANCE FOR TRACER 1 ***
INITIAL QUANTITY OF 1 : 0.000000 M3
FINAL QUANTITY : 0.6498000E-01 M3
QUANTITY ENTERED THROUGH LIQ. BND. : -0.9078299 M3
TOTAL MASS OF DEPOSIT : 11.47976 M3
TOTAL QUANTITY LOST : -12.45257 M3

Especially:
- QUANTITY ENTERED THROUGH LIQ. BND.: I can't relate the given (negative?) value to my input parameters
- TOTAL QUANTITY LOST: is it the quantity leaving the system through liquid boundaries? Or is it the lost mass according to mass balance check?

Finally, is it possible to define a sediment concentration inflow on a node inside the mesh, if possible with time varying concentrations defined in a source-like file? I did not find such a possibility when reading the user manual. This would be a great feature to simulate sediment plumes from dedging activities.

Thank you very much in advance for your help!

Best regards
PL

I also don't know it, But, the think in your sisyphe.file the 'CONCENTRATION PER CLASS AT BOUNDARIES' should be the same as the as the the sediment types.

In your case, you only have 0 type of class, so the value for CONCENTRATION PER CLASS AT BOUNDARIES also should be one value.

That is my personal idea

In your case, I also have a question about grain feeding. Did you set the CONCENTRATION PER CLASS AT BOUNDARIES by the trace set?

If you want to set the PRESCRIBED SOLID DISCHARGES, how could you set it ??

Hello,

"Normally" the suspended sediment equation should be mass-conservative and this has been checked a number of times. An important point is to choose a mass-conservative advection scheme for the suspended sediment, so the method of charecteristics is excluded, hence your keyword :

TYPE OF ADVECTION : 1

is the first thing to change, try 4 or 5 or 14 if you have tidal flats.

The advection equation for the suspended sediment is a very tricky issue because the advection field is corrected to account for the fact that the sediment is close to the bottom. The consequence is that the corrected advection field is not divergence free and a really conservative form of the equation must be solved. Another mathematical consequence is that monotonicity cannot be theoretically ensured, however this is controlled by deposition which tends to reduce high concentrations.

A "negative" quantity entered into the domain is an exiting quantity.

With best regards,

Jean-Michel Hervouet

Hello,

Thank you. Indeed, I use to have TYPE OF ADVECTION : 14 in my cases, but this one still had 1. Forgot to update it...
I am running a new simulation and will hopefully have better results.

Can you explain what is meant by the following terms in the log:

QUANTITY ENTERED THROUGH LIQ. BND.

and

ERROR ON VOLUME


Thank you in advance

Best regards
PL

Hello,

QUANTITY ENTERED THROUGH LIQ. BND.
is the:
quantity entered through liquid boundaries

The error on volume checks that final quantity - initial quantity - deposition = what entered in the domain.

REgards,

JMH